Again, because the interfaces that use filter-based forwarding must not be bound to a VPN, you configure an alternate method to provide next-hop routes to the VRF table by defining an interface routing table group and sharing this group among all the routing tables. To provide a route back to the clients for normal inet.0 routing, you define a static route to include in inet.0 and redistribute the static route into BGP:
- [edit]
- routing-options {
-
- interface-routes {
- rib-group inet if-rib;
- }
-
- rib-groups {
-
- if-rib {
- import-rib [ inet.0 vpn-B.inet.0 ];
- }
- }
- }
To direct traffic from VPN B to Client D, you configure the routing instance for VPN B on Router F. All incoming traffic from Client D on interface so-3/3/3.0 is forwarded normally by means of the destination address based on the routes in inet.0.
- [edit]
- routing-instances {
-
- vpn-B {
- instance-type vrf;
- route-distinguisher 172.21.10.64:200;
- vrf-import vpn-B-import;
- vrf-export vpn-B-export;
-
- routing-options {
-
- static {
- route 192.168.3.0/24 next-hop so-3/3/3.0;
- }
- }
- }
- }